The Metaphor Wars

Yesterday’s post was a response to the charge of incompleteness brought against the position that words are tools for piloting attention. The argument runs that attention cannot be directed toward unknown things; therefore, attention cannot be directed to ideas that are only imagined by the speaker. The listener cannot know what the speaker means.

This blog’s reply looked at metaphors. They direct the listener’s attention to some already-known feature of a new idea. They also provide a way of escaping the tautologies of logical systems; however, there are objections to claims for metaphors as the powerhouses of language.

Twenty six years ago, George Lakoff and Mark Johnson published a book titled Metaphors We Live By. It argued that metaphors are the tools by which people can “create a reality rather than simply to give way to the conceptualizing of a preexisting reality” (p. 144). Unfortunately, the book was published without an index, so I cannot tell quickly whether or not there were any incidental references to attention. But even without that, we can get the basic relationship: metaphors work by bringing to our attention experiences that are relevant but not part of the here and now.

Lakoff and Johnson gave more credit than seems necessary, at least for the purposes of this blog, to the notion of an underlying (and often unspoken) metaphorical concept. Thinking, for example, may use metaphors from physical handling—grasp an idea, turn over a suggestion, etc—without requiring a person to believe that an idea is literally a physical object that can be held, stolen, or turned on its head. And alternate metaphors are available. For example, ideas might be objects like stars that can be seen but not touched. He could sense the idea twinkling beyond the clouds, but could not quite make it out—a metaphor like that does not force a choice on its users. Which are they stars in the sky or sticks in the hand?

Lakoff says we can have both metaphorical concepts in our head, and that seems surely true since mixed metaphors are so common: Joe wrestled with Darwin’s ideas that whole semester until finally he grasped them, and from then on Darwinism became his pole star guiding him through biology's puzzling sea. Lakoff argues that we have both larger concepts stored in our brains, but that detail seems to add nothing to the analysis, especially since at any moment we can jump from one or the other, or be swayed by one or the other.

Lakoff is a former student of Chomsky’s who broke with him to study semantics rather than syntax. The split was recently reflected in a merciless review Stephen Pinker published in The New Republic (available here) that tore into Lakoff’s latest book, a treatise on the metaphors for freedom. (Lakoff’s reply is here) Ostensibly this review was about a political book, but the bitterness with which Pinker skewered Lakoff suggests something more was afoot.

Fortunately, Pinker does not keep his objection a secret. Metaphors are concrete while syntax is abstract. Pinker has bet on abstractions and he rejects the way Lakoff stresses the other side:

As many of Lakoff's skeptical colleagues have noted, the ubiquity of metaphor in language does not imply that all thinking is concrete. People cannot use a metaphor to reason with unless they have a deeper grasp of which aspects of the metaphor should be taken seriously and which should be ignored. When reasoning about a relationship as a kind of journey, it is fine to mull over the counterpart to a common destination, or to the bumpy stretches along the way -- but someone would be seriously deranged if he wondered whether he had time to pack, or whether the next gas station has clean restrooms. Thinking cannot trade in metaphors directly. It must use a more basic currency that captures the abstract concepts shared by the metaphor and its topic -- progress toward a shared goal in the case of journeys and relationships, conflict in the case of argument and war -- while sloughing off the irrelevant bits.

I have marked Pinker’s use of metaphor in this passage with a bold font. At first I thought Pinker might have been joking by using such a range of metaphors to say that at bottom we don’t really think metaphorically.

But then I noticed the rhetorical power of his mixed metaphor—(1) a more basic currency (2) that captures. Since real currencies don’t capture things, we might wonder why Pinker mixed the two. It turns out that mixing metaphors directs our attention away from a problem with Pinker's argument. The deeper confusion became apparent when I tried to make metaphors (1) and (2) more consistent. I quickly found that the controlling, earlier metaphor—trade in— was the real villain. I call it a villain because it was an unnecessary metaphor that does not get at some unknowable process. There was already a simple English verb available to Pinker: use, as in Thinking cannot use metaphors directly. And we can continue: It must use a more basic system that can manipulate the abstract concepts common to the metaphor and its topic while not using the irrelevant bits. But this sentence forces a new puzzle. Why bother with metaphors at all, if neither the speaker nor the listener uses them?

By introducing the “trades in” metaphor, Pinker has directed our attention away from questions like that and the mixed metaphor, captures, moves the reader even further from the objection. Stage magicians routinely shift attention, making the audience look to the left while something critical happens stage right, but Pinker was unlikely to have been so calculatedly devious. Probably, he, like the rest of us, really does think directly in metaphors and we don’t notice when we mislead ourselves and anybody else listening in on the conversation as we move the focus of our attention off the dime.

Metaphors are double edged. One side is well known to any logician; metaphors can confuse thinking quite hopelessly and should be kept out of any formal system for manipulating symbols. The other is well known to any writer or lover of literature; metaphors lie at the heart of language’s power to break the bonds of the intellectual here-and-now.

Comments

[Chris]

In that paragraph, Pinker is actually echoing an objection first voiced in the literature by Greg Murphy, and since quite commonly used against conceptual metaphor theory. The objection isn't that people can't use metaphors "directly," it's that unless the concepts involved in the metaphor already have some structure of their own, then you can't know what aspects of the base are relevant for structuring the target.

A good way to think about this is from Nelson Goodman's old criticism of similarity. In order to determine how similar two things are, Goodman argued, you have to know which properties are relevant. If you don't, then all things are infinitely similar to each other. An elephant and a supernova are infinitely similar, for example, because they're both larger than an ant, and larger than a butterfly, and larger than a blade of grass, and larger than the infinite number of things that are smaller than an elephant.

In metaphor, knowing which properties of the target and base are relevant to the comparison is even more important, because metaphors generally involve two very different concepts (very different on the surface, at least). Pinker's point, then, is that in order to figure out what properties of the base to carry over to the target, the target has to have structure of its own. Thus, the idea that abstract concepts are structured metaphorically becomes nonsensical.

This doesn't do anything to your point that metaphor (I would say analogy) guides attention. In fact, there's a ton of research on the role of analogy in things like attention and memory, and much of it shows that analogies do in fact guide our attention quite effectively. This is probably why so much of our language is metaphorical, even if the metaphors are lexicalized and long dead.

------------------------------
BLOGGER: The difference between an analogy and a metaphor is the same as the difference between a similie and a metaphor. One says A is like B and the other says A is B. When I say, "He grasped the idea," I am using a metaphor, I am not saying he did something akin to grasping.

[TLTB]

Pinker isn't arguing that metaphor has no role in thought. Rather, he is arguing against Lakoff's strong (and in my opinion unreasonable) position that the only thing we can think is metaphor. For Lakoff, metaphor and thought are nearly a tautology. Its unclear how critical thinking even arises in this sort of view, and indeed reading Lakoff's political essays one might think he is the only one gifted with that ability.

Pinker easily demonstrates that this cannot be the case; he also uncovers the real motives of Lakoff's political analysis in his discussion of motherhood vs. parenthood, showing that even if Lakoff is right about thought and metaphor, there is no empirical basis for the metaphors he chooses to indict.

I haven't read the book, but I have heard Lakoff give talks on its contents twice. In my opinion, Lakoff's only goal is to provide a (pseudo-)scientific justification for why liberals are better people than conservatives. It really is difficult to find any other motivation. It seems that even for Lakoff metaphors are tools, useful for manipulation but not the holistic basis for human thought.

----------------------
BLOGGER: i agree that Lakoff has done neither his theory nor liberalism any favors by thring to use his notions of metaphors to provide political advice. Lakoff's application shows the weakness of his interest in an unspoken metaphorical concept, e.g., conservatives think of government as a father, liberals as a mother.

I'm not as interested in Lakoff's theory as I am in Pinker's attack. Pinker is not rejecting the (ridiculuous) notion that all thinking is metaphorical. He is denying that "all thinking is concrete<' and then goes on to imply that ultimately all thinking is abstract, "cannot trade in metaphors directly."

I do not argue that all thinking is concrete, but that at least some of it is and that metaphors provide a way of thinking concretely outside the here and now.

I'm not sure if Chris's comment voices opposition to Lakoff's attention to the unspoken metaphorical concept or to the idea of metaphorical and concrete thinking in general. If the former, I'm with him; if the latter, I'm not.

[Giorgio Marchetti]

How to direct attention to ideas and thoughts that are only imagined by the speaker, and that the listener cannot know? This is really one of the hardest problem in language acquisition studies (and, as far as I understand now thanks to the discussion that is goin on in this wonderful blog, also in language and thought evolution studies).

Certainly, metaphors can explain something: but only as long as the listener alredy knows something about the target and the base. The "Metaphor mechanism" seems to me to hold only in adults or in children who already have a certain knowledge of their world.

Very clear examples of something that is only in the speaker's head and that the listener cannot know are supplied by all those instances in which a mother tries to explain to her young child all those abstract realtions between things, objects and so on that are usually expressed by means of prepositions, conjunctions and so on. These relations are not in the world of "real things". While a "car" is visible, is "there", the relation between this car and aaddy is not "there": "look at daddy's car!".

I do not think that metaphors can explain that "'s" (at least as far as babies are concerned). A mother uses gestures, pilotes her baby's attention primarily with her hands, eyes, arms, the tone of the voice, and so on.

Language learning (and also evolution, I believe) starts with perceived things, gestures, actions (which only can pilot attention, at the very beginning of the process of acquisition). Metaphors come later.

Giorgio Marchetti